Cees Snoek

What defines an event in video? Answers from the recent literature indicate success can be obtained with a color Fisher vector, a histogram of motion and trajectories, or, even better, a potpourri of multimedia descriptors and representations. In this talk I will highlight our progress on encoding video, and events, by semantic detector predictions, which can not only recognize but also explain events. First I will present our study on the characteristics of a universal semantic encoding for arbitrary-event recognition in web video. Then I will introduce an algorithm that learns from examples what concepts in a semantic encoding are most informative per event. Finally, I will end by showing event recounting capabilities of the semantic encodings, which open up the possibility to automatically describe and explain why a particular video was found.

Cees G. M. Snoek is currently an associate professor at the University of Amsterdam. He was previously at Carnegie Mellon University and the University of California at Berkeley. His research interest is video and image search. Dr. Snoek is the principal investigator of the MediaMill Semantic Video Search Engine, which is a consistent top performer in the yearly NIST TRECVID evaluations. He is member of the editorial boards for IEEE Multimedia and IEEE Transactions on Multimedia, and general co-chair of ACM Multimedia 2016. Cees is recipient of an NWO Veni award, an NWO Vidi award, and the Netherlands Prize for ICT Research 2012. Several of his Ph.D. students have won best paper awards, including the IEEE Transactions on Multimedia Prize Paper Award and the SIGMM Best Ph.D. Thesis Award.